feat: add configurable LLM settings for mem0_memory tool (#221)

- Add support for MEM0_LLM_PROVIDER, MEM0_LLM_MODEL, MEM0_LLM_TEMPERATURE, MEM0_LLM_MAX_TOKENS environment variables
- Add support for MEM0_EMBEDDER_PROVIDER, MEM0_EMBEDDER_MODEL environment variables
- Fix DEFAULT_CONFIG to properly use os.environ.get() with defaults
- Add comprehensive tests for environment variable configuration
- Update README.md documentation with new configuration options
- Add module docstring documentation for new environment variables

Fixes #220

Co-authored-by: Hemanth <[email protected]>

Author: ChallapalliHemanthsaikumar

README.md

@@ -759,12 +759,19 @@ The Mem0 Memory Tool supports three different backend configurations:
 | OPENSEARCH_HOST | OpenSearch Host URL | None | OpenSearch |
 | AWS_REGION | AWS Region for OpenSearch | us-west-2 | OpenSearch |
 | DEV | Enable development mode (bypasses confirmations) | false | All modes |
+| MEM0_LLM_PROVIDER | LLM provider for memory processing | aws_bedrock | All modes |
+| MEM0_LLM_MODEL | LLM model for memory processing | anthropic.claude-3-5-haiku-20241022-v1:0 | All modes |
+| MEM0_LLM_TEMPERATURE | LLM temperature (0.0-2.0) | 0.1 | All modes |
+| MEM0_LLM_MAX_TOKENS | LLM maximum tokens | 2000 | All modes |
+| MEM0_EMBEDDER_PROVIDER | Embedder provider for vector embeddings | aws_bedrock | All modes |
+| MEM0_EMBEDDER_MODEL | Embedder model for vector embeddings | amazon.titan-embed-text-v2:0 | All modes |
+
 
 **Note**:
 - If `MEM0_API_KEY` is set, the tool will use the Mem0 Platform
 - If `OPENSEARCH_HOST` is set, the tool will use OpenSearch
 - If neither is set, the tool will default to FAISS (requires `faiss-cpu` package)
-
+- LLM configuration applies to all backend modes and allows customization of the language model used for memory processing
 #### Memory Tool
 
 | Environment Variable | Description | Default |

src/strands_tools/mem0_memory.py

@@ -150,13 +150,16 @@ class Mem0ServiceClient:
     """Client for interacting with Mem0 service."""
 
     DEFAULT_CONFIG = {
-        "embedder": {"provider": "aws_bedrock", "config": {"model": "amazon.titan-embed-text-v2:0"}},
+        "embedder": {
+            "provider": os.environ.get("MEM0_EMBEDDER_PROVIDER", "aws_bedrock"),
+            "config": {"model": os.environ.get("MEM0_EMBEDDER_MODEL", "amazon.titan-embed-text-v2:0")},
+        },
         "llm": {
-            "provider": "aws_bedrock",
+            "provider": os.environ.get("MEM0_LLM_PROVIDER", "aws_bedrock"),
             "config": {
-                "model": "anthropic.claude-3-5-haiku-20241022-v1:0",
-                "temperature": 0.1,
-                "max_tokens": 2000,
+                "model": os.environ.get("MEM0_LLM_MODEL", "anthropic.claude-3-5-haiku-20241022-v1:0"),
+                "temperature": float(os.environ.get("MEM0_LLM_TEMPERATURE", 0.1)),
+                "max_tokens": int(os.environ.get("MEM0_LLM_MAX_TOKENS", 2000)),
             },
         },
         "vector_store": {

tests/test_mem0.py

@@ -54,7 +54,18 @@ def extract_result_text(result):
     return str(result)
 
 
-@patch.dict(os.environ, {"OPENSEARCH_HOST": "test.opensearch.amazonaws.com"})
+@patch.dict(
+    os.environ,
+    {
+        "MEM0_LLM_PROVIDER": "openai",
+        "MEM0_LLM_MODEL": "gpt-4o",
+        "MEM0_LLM_TEMPERATURE": "0.2",
+        "MEM0_LLM_MAX_TOKENS": "4000",
+        "MEM0_EMBEDDER_PROVIDER": "openai",
+        "MEM0_EMBEDDER_MODEL": "text-embedding-3-large",
+        "OPENSEARCH_HOST": "test.opensearch.amazonaws.com",
+    },
+)
 @patch("strands_tools.mem0_memory.Mem0Memory")
 @patch("strands_tools.mem0_memory.boto3.Session")
 def test_store_memory(mock_boto3_session, mock_mem0_memory, mock_tool):